Meta-analysis of randomized controlled trials comparing the risk of postoperative infection between recipients of allogeneic and autologous blood transfusion

Background and Objectives A previous meta‐analysis of randomized controlled trials (RCTs), comparing the risk of postoperative infection between recipients of allogeneic blood or autologous blood obtained by preoperative autologous blood donation (PABD), did not detect an immunomodulatory (TRIM) effect of allogeneic blood transfusion (ABT). If such a TRIM effect was mediated by white blood cell (WBC)‐derived soluble mediators accumulating during storage, however, stored autologous blood obtained by PABD would not prevent the TRIM effect, whereas unstored autologous blood obtained by acute normovolemic haemodilution (ANH), intraoperative blood recovery (IBR), or postoperative blood recovery (PBR), would abrogate the TRIM effect. Materials and Methods RCTs reported through January 2002 were retrieved, and five studies met the criteria for meta‐analysis. Summary odds ratios (ORs) of postoperative infection in recipients of allogeneic vs. autologous blood were calculated across studies. Results No difference in the risk of infection between the comparison arms was detected across all five RCTs [summary OR = 1·22, 95% confidence interval (95% CI): 0·75–1·98], or when the results of studies using PABD or ANH/IBR/PBR were integrated separately (summary OR = 1·35; 95% CI: 0·45–4·08; and summary OR = 1·49; 95% CI: 0·69–3·22, respectively). Conclusions The finding of no TRIM effect of ABT across RCTs using ANH/IBR/PBR to obtain autologous blood does not support the hypothesis that a TRIM effect of ABT is mediated by WBC‐derived soluble mediators accumulating during storage. The null finding of the overall meta‐analysis also does not support a TRIM effect of ABT mediated by other blood constituents.